The Functional Assessment of Cancer Therapy Scale: Development and Validation of the General Measure

PURPOSE

We developed and validated a brief, yet sensitive, 33-item general cancer quality-of-life (QL) measure for evaluating patients receiving cancer treatment, called the Functional Assessment of Cancer Therapy (FACT) scale.

METHODS AND RESULTS

The five-phase validation process involved 854 patients with cancer and 15 oncology specialists. The initial pool of 370 overlapping items for breast, lung, and colorectal cancer was generated by open-ended interview with patients experienced with the symptoms of cancer and oncology professionals. Using preselected criteria, items were reduced to a 38-item general version. Factor and scaling analyses of these 38 items on 545 patients with mixed cancer diagnoses resulted in the 28-item FACT-general (FACT-G, version 2). In addition to a total score, this version produces subscale scores for physical, functional, social, and emotional well-being, as well as satisfaction with the treatment relationship. Coefficients of reliability and validity were uniformly high. The scale's ability to discriminate patients on the basis of stage of disease, performance status rating (PSR), and hospitalization status supports its sensitivity. It has also demonstrated sensitivity to change over time. Finally, the validity of measuring separate areas, or dimensions, of QL was supported by the differential responsiveness of subscales when applied to groups known to differ along the dimensions of physical, functional, social, and emotional well-being.

CONCLUSION

The FACT-G meets or exceeds all requirements for use in oncology clinical trials, including ease of administration, brevity, reliability, validity, and responsiveness to clinical change. Selecting it for a clinical trial adds the capability to assess the relative weight of various aspects of QL from the patient's perspective.

Validation and clinical utility of the executive function performance test in persons with traumatic brain injury

This study examined the relationships between the Executive Function Performance Test (EFPT), the NIH Toolbox Cognitive Function tests, and neuropsychological executive function measures in 182 persons with traumatic brain injury (TBI) and 46 controls to evaluate construct, discriminant, and predictive validity. Construct validity: There were moderate correlations between the EFPT and the NIH Toolbox Crystallized (r = -.479), Fluid Tests (r = -.420), and Total Composite Scores (r = -.496). Discriminant validity: Significant differences were found in the EFPT total and sequence scores across control, complicated mild/moderate, and severe TBI groups. We found differences in the organisation score between control and severe, and between mild and severe TBI groups. Both TBI groups had significantly lower scores in safety and judgement than controls. Compared to the controls, the severe TBI group demonstrated significantly lower performance on all instrumental activities of daily living (IADL) tasks. Compared to the mild TBI group, the controls performed better on the medication task, the severe TBI group performed worse in the cooking and telephone tasks. Predictive validity: The EFPT predicted the self-perception of independence measured by the TBI-QOL (beta = -0.49, p < .001) for the severe TBI group. Overall, these data support the validity of the EFPT for use in individuals with TBI.

Outcome measures in spinal cord injury: recent assessments and recommendations for future directions

STUDY DESIGN

Review by the spinal cord outcomes partnership endeavor (SCOPE), which is a broad-based international consortium of scientists and clinical researchers representing academic institutions, industry, government agencies, not-for-profit organizations and foundations.

OBJECTIVES

Assessment of current and evolving tools for evaluating human spinal cord injury (SCI) outcomes for both clinical diagnosis and clinical research studies.

METHODS

a framework for the appraisal of evidence of metric properties was used to examine outcome tools or tests for accuracy, sensitivity, reliability and validity for human SCI.

RESULTS

Imaging, neurological, functional, autonomic, sexual health, bladder/bowel, pain and psychosocial tools were evaluated. Several specific tools for human SCI studies have or are being developed to allow the more accurate determination for a clinically meaningful benefit (improvement in functional outcome or quality of life) being achieved as a result of a therapeutic intervention.

CONCLUSION

Significant progress has been made, but further validation studies are required to identify the most appropriate tools for specific targets in a human SCI study or clinical trial.

Development of a general ability index for the Wechsler Adult Intelligence Scale--Third Edition

The Wechsler Adult Intelligence Scale--Third Edition (WAIS-III; D. Wechsler, 1997) permits the calculation of both traditional IQ and index scores. However, if only the subtests constituting the index scores are administered, especially those yielding the Verbal Comprehension and Perceptual Organization Indexes, there is no equivalent measure of Full Scale IQ. Following the procedure for calculating a General Ability Index (GAI; A. Prifitera, L. G. Weiss, & D. H. Saklofske, 1998) for the Wechsler Intelligence Scale for Children--Third Edition (D. Wechsler, 1991), GAI normative tables for the WAIS-III standardization sample are reported here.

Development of a general ability index for the Wechsler Adult Intelligence Scale--Third Edition

The Wechsler Adult Intelligence Scale--Third Edition (WAIS-III; D. Wechsler, 1997) permits the calculation of both traditional IQ and index scores. However, if only the subtests constituting the index scores are administered, especially those yielding the Verbal Comprehension and Perceptual Organization Indexes, there is no equivalent measure of Full Scale IQ. Following the procedure for calculating a General Ability Index (GAI; A. Prifitera, L. G. Weiss, & D. H. Saklofske, 1998) for the Wechsler Intelligence Scale for Children--Third Edition (D. Wechsler, 1991), GAI normative tables for the WAIS-III standardization sample are reported here.

WAIS-III reliability data for clinical groups

Reliability estimates for psychological tests are almost always reported for nonclinical populations (e.g., the normative samples). Such practice will no longer be sufficient as the new standards for testing call for an adequate assessment of psychometric properties within the specific population being tested. The purpose of this study was to provide internal consistency reliability estimates for clinical groups on the Wechsler Adult Intelligence Scale-Third Edition. The study included data from 403 clinical participants composed of 10 groups of adults recruited as part of the WAIS-III clinical validity studies. Split-half reliability coefficients were obtained for these groups replicating the procedure used in the WAIS-III. With 8 of the clinical groups, the split-half reliability coefficients were comparable to, or even higher than, those reported for the WAIS-III standardization sample. In general, the split-half coefficients for the Verbal subtests tended to be higher than the coefficients for the Performance subtests. The high magnitude and general pattern of these coefficients demonstrate that the WAIS-III scales do not include additional error variance above and beyond what is reported in the WAIS-III-WMS-III Technical Manual when it was used to assess certain clinical groups. For the ADHD/ADD and learning disabilities groups, however, the internal consistencies coefficients of some subtests were relatively lower, although not statistically significant, than the normative sample. These findings may reflect more heterogeneity within the groups. The implications for assessment and for using alternate methods of determining the psychometric properties in these populations are discussed.

The influence of IQ stratification on WAIS-III/WMS-III FSIQ-general memory index discrepancy base-rates in the standardization sample

Since memory performance expectations may be IQ-based, unidirectional base rate data for IQ-Memory Score discrepancies are provided in the WAIS-III/WMS-III Technical Manual. The utility of these data partially rests on the assumption that discrepancy base rates do not vary across ability levels. FSIQ stratified base rate data generated from the standardization sample, however, demonstrate substantial variability across the IQ spectrum. A superiority of memory score over FSIQ is typical at lower IQ levels, whereas the converse is true at higher IQ levels. These data indicate that the use of IQ-memory score unstratified "simple difference" tables could lead to erroneous conclusions for clients with low or high IQ. IQ stratified standardization base rate data are provided as a complement to the "predicted difference" method detailed in the Technical Manual.

Criterion validity of new WAIS-II subtest scores after traumatic brain injury

The criterion validity of the new subtests from the Wechsler Adult Intelligence Scale-Third Edition (WAIS-III; Wechsler, 1997) was evaluated in a sample of 100 patients with traumatic brain injury (TBI). Letter-Number Sequencing and Symbol Search, but not Matrix Reasoning, yielded statistically significant differences in performance between patients with moderate-severe TBI, patients with mild TBI, and demographically matched controls. Level of education accounted for a statistically significant amount of variance in the performance of patients with TBI, in addition to that explained by injury severity variables. It is concluded that Letter-Number Sequencing and Symbol Search have satisfactory criterion validity, but that they need to be supplemented with other measures in the context of neuropsychological evaluations. Matrix Reasoning, on the other hand, is not sensitive to the sequelae of TBI and more studies are needed to determine how it can be used for neuropsychological assessment purposes.

WAIS-III and WMS-III profiles of mildly to severely brain-injured patients

Wechsler Adult Intelligence Scale-III (WAIS-III) and Wechsler Memory Scale-III (WMS-III; The Psychological Corporation, 1997) scores of patients with mild traumatic brain injury (MTBI, n = 23) to moderate-severe traumatic brain injury (M-S TBI, n = 22) were compared to those of 45 matched normal control patients. WAIS-III results revealed that IQ and index scores of MTBI patients did not significantly differ from those of controls, whereas M-S TBI patients received significantly lower mean scores on all measures. All M-S TBI patients' WMS-III index scores also revealed significantly lower scores in comparison to those of control participants, with the exception of Delayed Auditory Recognition. MTBI patients showed significantly lower mean index scores compared to normal controls on measures of immediate and delayed auditory memory, immediate memory, visual delayed memory, and general memory. Eta-squared analyses revealed that WMS-III visual indexes and WAIS-III processing speed showed particularly large effect sizes. These results suggest that symptomatic MTBI patients obtain some low WMS-III test scores comparable to those of more severely injured patients.

Two-tailed versus one-tailed base rates of discrepancy scores in the WAIS-III

Tables included in the WAIS-III report the frequency of discrepancies between IQ and index scores independent of the directionality of the score. If an examiner does not have any a priori hypothesis about which skill may be more developed and which skills may be weaknesses, it is appropriate to use these tables as they are. In this case the examiner would be looking for unusually large discrepancies between scores irrespective of direction. When an examiner has a hypothesis about which skills may be weaknesses and strengths for an individual, the frequencies that are based upon an absolute value of the discrepancy will cause examiners to overestimate the frequency of the occurrence of the discrepancy score in question. Sattler and Ryan (1998) and Tulsky, Zhu, and Vasquez (1998) suggested dividing the frequencies reported in the WAIS-III tables in half to obtain the correct base rate. This suggestion is tested in this paper. New observed frequency tables were derived from the WAIS-III and WMS-III standardization samples and these frequencies were compared against the estimated frequencies using the method described by Sattler and Ryan (1998) and Tulsky et al. (1998). The differences between these two methods were calculated and are, for the most part, insignificant. In light of the similarity between the methods, the implications of using observed frequencies versus estimated frequencies is discussed.

Could test length or order affect scores on letter number sequencing of the WAIS-III and WMS-III? Ruling out effects of fatigue

The Letter Number Sequencing subtest of the WAIS-III and WMS-III was administered at the end of the standardization edition of the WMS-III. It was not administered as part of the WAIS-III standardization battery. Nevertheless, the subtest was included in the published version of the WAIS-III. This study examines differences between examinees administered the Letter Number Sequencing subtest at three different times during a psychological battery: (1) as part of the published battery, (2) as part of the WMS-III when the WMS-III was administered as the first test in a sequence, and (3) as part of the WMS-III standardization when the WAIS-III was administered immediately preceding the WMS-III. The participants were 372 examinees ( n = 124 in each condition) who were matched on key demographic variables. A repeated measures MANOVA yielded no difference in subtest scores when administered in any of these conditions. The results show no evidence of fatigue or ordering effects on the Letter Number Sequencing subtest.

Co-norming the WAIS-III and WMS-III: Is there a test-order effect on IQ and memory scores?

Test-order effect on the WAIS-III and WMS-III scores was evaluated using the WMS-III standardization sample. Participants completed the standardization editions of the WAIS-III and WMS-III in one session, with the tests administered in roughly counterbalanced order. Repeated measure MANOVA analyses were conducted to determine if there was an overall test-order effect for subtest, index, or IQ scores. No significant test-order effects were found for either the WAIS-III index or IQ scores or for the WMS-III index scores. At the subtest level, the majority of the WAIS-III and WMS-III subtests did not show a significant test-order effect. The exceptions were Digit Span and Digit Symbol-Coding on the WAIS-III and Faces II and Logical Memory II on the WMS-III. Although statistically significant test-order effects were found on these subtests, the effect sizes were small. This study indicates that the test-order effect is not a potential threat to the internal validity of the WAIS-III and WMS-III normative data. The practical implications of the current study are discussed.

Reliability and validity of the Functional Assessment of Cancer Therapy-Colorectal (FACT-C) quality of life instrument

BACKGROUND

This paper describes the development of the Functional Assessment of Cancer Therapy-Colorectal (FACT-C)--a questionnaire assessing quality of life concerns pertinent to colorectal cancer patients. This self-report instrument combines specific concerns related to colorectal cancer with concerns that are common to all cancer patients as assessed with the FACT-General (FACT-G).

METHOD

Three separate and distinct validation samples were used with different ethnic and stage of disease compositions.

RESULTS

Results suggest that the FACT-C is a reliable and valid measure in both its English and Spanish language versions. Specifically, the FACT-C evidenced good internal consistency reliability and concurrent validity, as well as an ability to distinguish between groups based on functional status and extent of disease. The FACT-C was also found to be sensitive to changes in functional status. Colorectal cancer-specific items form the Colorectal Cancer Subscale (CCS) which was found to have adequate convergent and divergent validity. Internal consistency was adequate in most samples. In addition, the CCS was able to distinguish among groups that differed in functional status and was sensitive to change in functional status.

CONCLUSION

It is recommended that the entire FACT-C (rather than simply the CCS) be used when studying patients with colorectal cancer in order to provide a comprehensive assessment of quality of life. Recommendations are offered concerning the appropriate use of these measures in clinical research and directions for future research with colorectal cancer patients.

Reliability and validity of the Functional Assessment of Cancer Therapy-Breast quality-of-life instrument

PURPOSE

This is the first published report on the validation of the Functional Assessment of Cancer Therapy-Breast (FACT-B), a 44-item self-report instrument designed to measure multidimensional quality of life (QL) in patients with breast cancer. The FACT-B consists of the FACT-General (FACT-G) plus the Breast Cancer Subscale (BCS), which complements the general scale with items specific to QL in breast cancer. The FACT-B was developed with an emphasis on patients' values and brevity and is available in nine languages.

METHODS AND RESULTS

Two validation samples were used for this report. The first (n = 47) was tested twice over a 2-month period to assess sensitivity to change. Significant sensitivity to change in performance status rating (PSR) was demonstrated for the FACT-B total score, the Physical Well-Being (PWB) subscale, the Functional Well-Being (FWB) subscale, and the BCS. Sensitivity to change in QL as measured by the Functional Living Index-Cancer (FLIC) was documented in the FACT-B total score, PWB, FWB, and Emotional Well-Being (EWB). Additional validity and reliability data were obtained from a larger sample (n = 295). The alpha coefficient (internal consistency) for the FACT-B total score was high (alpha = .90), with subscale alpha coefficients ranging from .63 to .86. Evidence supported test-retest reliability, as well as convergent, divergent, and known groups validity.

CONCLUSION

The FACT-B is appropriate for use in oncology clinical trials, as well as in clinical practice. It demonstrates ease of administration, brevity, reliability, validity, and sensitivity to change.

Reliability and validity of the Functional Assessment of Cancer Therapy-Lung (FACT-L) quality of life instrument

The FACT-L (version 3) is a 44-item self-report instrument which measures multidimensional quality of life. Available in eight languages, it is currently being used in several Phase II and III lung cancer clinical trials. Reliability and validity of the 33-item version 2 of the FACT-General (FACT-G) have previously been published. This paper reports further validation data on the FACT-G with a subsample of lung cancer patients from the original publication and, more importantly, presents data on the Lung Cancer Subscale (LCS). The nine LCS questions were administered along with the FACT-G to 116 patients with lung cancer. Internal consistency (coefficient alpha) was improved from 0.53 to 0.68 by dropping two questions which were uncorrelated with the others. A subset of 41 patients was tested again at 2 months to evaluate sensitivity to change in performance status rating (PSR) and to obtain estimates of a clinically meaningful change score for the FACT-G and the 7-item LCS. Using a linear test for trend, sensitivity to change in performance status rating (PSR) was obtained with the Total score (P = 0.03), the Physical Well Being (PWB) subscale (P = 0.02), the Functional Well Being (FWB) subscale (P = 0.05), and the LCS (P = 0.03). A 21-item Trial Outcome Index (TOI), combining scores on PWB, FWB and LCS, was highly reliable (coefficient a = 0.89) and sensitive to change in PSR F(1,38) = 4.84 (P = 0.01). This TOI is probably the most relevant and precise indicator of patient-reported quality of life available for lung cancer patients who complete the FACT-L while participating in an oncology clinical trial. The FACT-L may also be of benefit in evaluating quality of life in patients with lung diseases other than cancer.

The Functional Assessment of Cancer Therapy scale: development and validation of the general measure

PURPOSE

We developed and validated a brief, yet sensitive, 33-item general cancer quality-of-life (QL) measure for evaluating patients receiving cancer treatment, called the Functional Assessment of Cancer Therapy (FACT) scale.

METHODS AND RESULTS

The five-phase validation process involved 854 patients with cancer and 15 oncology specialists. The initial pool of 370 overlapping items for breast, lung, and colorectal cancer was generated by open-ended interview with patients experienced with the symptoms of cancer and oncology professionals. Using preselected criteria, items were reduced to a 38-item general version. Factor and scaling analyses of these 38 items on 545 patients with mixed cancer diagnoses resulted in the 28-item FACT-general (FACT-G, version 2). In addition to a total score, this version produces subscale scores for physical, functional, social, and emotional well-being, as well as satisfaction with the treatment relationship. Coefficients of reliability and validity were uniformly high. The scale's ability to discriminate patients on the basis of stage of disease, performance status rating (PSR), and hospitalization status supports its sensitivity. It has also demonstrated sensitivity to change over time. Finally, the validity of measuring separate areas, or dimensions, of QL was supported by the differential responsiveness of subscales when applied to groups known to differ along the dimensions of physical, functional, social, and emotional well-being.

CONCLUSION

The FACT-G meets or exceeds all requirements for use in oncology clinical trials, including ease of administration, brevity, reliability, validity, and responsiveness to clinical change. Selecting it for a clinical trial adds the capability to assess the relative weight of various aspects of QL from the patient's perspective.

Quality of life in cancer: definition, purpose, and method of measurement

Over the past 20 years, changes within medicine and society have encouraged an expansion in conceptualizing and evaluating the goals and important outcomes of medical treatments. In evaluating the care of people with cancer and other chronic illnesses, a major part of the expansion has been the inclusion of means to assess the quality of life in addition to the quantity of life associated with various treatments. This paper presents a review of the current literature on quality-of-life assessment in oncology research and practice. The review is organized around current thinking about the definition of quality of life, the purpose of quality-of-life measurement, the method of assessment, and the competing advantages and disadvantages of disease-specific versus general health outcome measurement. Recommendations for proceeding with quality-of-life measurement in oncology are provided to encourage increased attention to the need to remain closely tied to a theoretical construct.

An introduction to test theory

A basic understanding of test theory is an important prerequisite before one may create or use a scale to measure behavior. This paper provides a basic review of the tenets of measurement theory. Descriptions of the different types of reliability (i.e., test-retest, parallel forms, split-halves, internal consistency) and validity (content, criterion-related, and construct) are defined and explained. Suggestions for test construction are advanced and techniques of generating and then narrowing a large pool of preliminary items are discussed. Finally, specific questions that should be asked before a test is used in a clinical or research setting are proposed.

Measuring quality of life today: methodological aspects

While it is recognized that quality of life is ultimately as important as quantity of life, efforts to implement quality of life measurement often fail. Two basic reasons for that failure include: 1) definitional differences, where different investigators attribute different meaning to the term and, as a result, are measuring different endpoints; and 2) insufficient information about available measures, which can lead to improper test selection and unnecessary regeneration of new items. Included in this paper is a table that reviews many available quality of life measures that have been designed for, or frequently used with, people with cancer. Proper selection of measures and supplementary questions is an important first step toward a successful evaluation of quality of life.